Multi-agent System Architectural Aspects for Continuous Replanning

Abstract

Robots’ coordination to achieve the system’s goal is one of the challenges that complex Multi-Robot Systems (MRS) encounter. One could use automated planning (AP) to better face this challenge by diminishing problems and continually correcting the execution when failures occur. Some works in the literature try to fix this problem, but there are still few, and there’s not much analysis between them. This work implements a Multi-Agent System (MAS) to simulate an MRS mission using a MAS architecture integrated with AP illustrated with space resource gathering robots. The results show the importance of the ability to plan recovery and research in complex space missions field.